You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Steam Controller can be used in two modes basically:
Mouse/keyboard emulation going through libinput (check libnput --list-devices)
Higher-level logic internal to the Steam-client.
Current issues:
In mode 1
Works in X11 sessions, but not in desktop Wayland sessions at the moment, tested on KWin Wayland, see for GNOME [1] and [2]. Apparently has worked in the past though (sure?). Interestingly a short test showed that it works in sway. Maybe the other compositors just don't support multiple pointer devices with different characteristic (i.e. different device groups etc.)?
In mode 2
Is activated once the Steam client starts up. Mode 1 gets deactivated at that point in time. Has some more gimmicks. But it uses XTestFakeRelativeMotionEvent for sending pointer motions to the OS. Changing the cursor position by that is currently not supported in XWayland (warping pointers is).
So even if problem of mode 1 is solved as soon as mode 2 is activated, i.e. the Steam compositor starts up, the pointer input should be broken in any current Wayland desktop session again.
In gamescope currently the compositor itself listens into the X11 connection to get the XTest induced cursor position. That blocks us from going Wayland-determined cursor-position only at the moment and leads to corner-case issues when the synchronization lags behind (mouse position only updated after first move, returns to last position where it was moved with mouse before controller etc.).
Solution ideas
Without changing the Steam-client internal code solutions could be:
For (2): support XTestFakeRelativeMotionEvent in XWayland.
...
The text was updated successfully, but these errors were encountered:
Steam Controller can be used in two modes basically:
libnput --list-devices
)Current issues:
In mode 1
Works in X11 sessions, but not in desktop Wayland sessions at the moment, tested on KWin Wayland, see for GNOME [1] and [2]. Apparently has worked in the past though (sure?). Interestingly a short test showed that it works in sway. Maybe the other compositors just don't support multiple pointer devices with different characteristic (i.e. different device groups etc.)?
In mode 2
Is activated once the Steam client starts up. Mode 1 gets deactivated at that point in time. Has some more gimmicks. But it uses XTestFakeRelativeMotionEvent for sending pointer motions to the OS. Changing the cursor position by that is currently not supported in XWayland (warping pointers is).
So even if problem of mode 1 is solved as soon as mode 2 is activated, i.e. the Steam compositor starts up, the pointer input should be broken in any current Wayland desktop session again.
In gamescope currently the compositor itself listens into the X11 connection to get the XTest induced cursor position. That blocks us from going Wayland-determined cursor-position only at the moment and leads to corner-case issues when the synchronization lags behind (mouse position only updated after first move, returns to last position where it was moved with mouse before controller etc.).
Solution ideas
Without changing the Steam-client internal code solutions could be:
The text was updated successfully, but these errors were encountered: